Course objectives
This is a course designed to introduce students to the use of mixed effects models with application using real biological data. Each week, there will be a 50-minute lecture on the topic of the week, followed a couple of days later by a 2-hour workshop, where we will explore a topic by working through assigned homework using the statistical program, R.
Students taking this course will develop advanced skills in using R for analyzing and graphing biological data. Specifically, students will learn advanced techniques for analysing data when there is a mixture of fixed and random effects. Emphasis will be placed on when and how to include random effects in statistical models as well as a practical understanding of the underlying statistical theory.
Instructor
Chris Eckert
When and where
- Tuesdays 100-230pm BioSciences rm 3110
- Thursdays 300-430pm BioSciences rm 3112
Helpful books
- Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM. 2009. Mixed effect models and extensions in ecology with R. New York, New York, USA: Springer.
- Faraway JJ. 2006. Extending the linear model with R. Boca Raton, Florida, USA: Chapman & Hall/CRC.
- Crawley MJ. 2007. The R book. Chichester, West Sussex, England: John Wiley and Sons.
Evaluation
- 20% Participation
- 40% Homework assignments (four @ 10% each)
- 40% Major assignment (due before 5pm Friday 19 February 2016)
Course schedule
Date
|
Topic
|
Week 1
|
Lecture 1: What are random effects and why we like them?
Workshop 1: Fixed effects versus random effects
|
Week 2
|
Lecture 2: Random intercepts and slopes models
|
Workshop 2: Running a simple mixed effects model and testing hypotheses
|
|
Week 3
|
Lecture 3: Simulating data with random effects.
Assignment 1 handed out |
Workshop 3: Nested models and variance components
Assignment 2 handed out
|
|
Week 4
|
Lecture 4: Repeated-measures designs
Assignment 1 due |
Workshop 4: Repeated-measures case study
Assignment 2 due (take up in class)
Assignment 3 handed out
|
|
Week 5
|
Lecture 5: Partially hierarchical designs
|
Workshop 5: Evaluating significance in LMMs
Assignment 3 due (take up in class)
Assignment 4 handed out
|
|
Week 6
|
Lecture 6: Multiple comparisons and mixed models
|
Workshop 6: Presenting results of mixed models
Assignment 4 due (take up in class)
|
|
Week 7
| Reading Week Fri 19 Feb, 5pm - Major Assignment Due (40%) |
The main goal of this course is to get you using mixed-effects statistical models in your own research. To that end, the major assignment is to use what you’ve learned in this course to analyze your own data: from your current thesis research, your undergraduate thesis or, barring that, a dataset somewhat related to your research. The only criterion is that the experiment or study or empirical survey must involve both random and fixed effects so that mixed-effects modelling is appropriate.
Here’s what I’d like your final assignment to consist of:
(1) A very, very brief introduction to the hypotheses being investigated, culminating in a clearly stated set of empirical predictions that you can and will test with you mixed models.
(2) A paragraph or two summarizing briefly how the data were collected to test your hypotheses, followed by a complete description of the statistical analysis (as in a scientific journal article). Make it clear what hypothesis each facet of the analysis addresses. I recommend reading the statistical part of the methods sections in relevant journal articles to see how other scientists write this section.
(3) A results section as in a scientific paper. This is a section that I will pay particular attention to. Writing about data and results is a fine craft and I encourage you to read relevant papers to get some ideas on what to say and what to avoid. Make sure you reference your figures and tables properly and try to avoid redundancy between the text and data displays.
(4) Figures and tables (as necessary) summarizing and presenting the results. Figures and tables serve different purposes, so use them appropriately.
(5) A very brief discussion section interpreting the results. I’m look for only one or two paragraphs here.
Please email the file to me as one PDF with tables and then figures embedded at the end of the text. Please write in the active voice throughout, it’s so much more efficient and engaging than the passive voice. Don’t forget to proof-read your paper carefully, run your spell-checker, and check THIS LIST to make sure you haven’t used any words or expression that I hate.
The paper will be marked out of 40, which includes marks for style, clarity, economy, grammar and spelling.
